[llvm-dev] LLVM Weekly - #280, May 13th 2019

Alex Bradbury via llvm-dev llvm-dev at lists.llvm.org
Mon May 13 13:17:06 PDT 2019

LLVM Weekly - #280, May 13th 2019

If you prefer, you can read a HTML version of this email at

Welcome to the two hundred and eightieth issue of LLVM Weekly, a weekly
newsletter (published every Monday) covering developments in LLVM, Clang, and
related projects. LLVM Weekly is brought to you by [Alex
Bradbury](https://www.linkedin.com/in/alex-bradbury/). Subscribe to future
issues at <http://llvmweekly.org> and pass it on to anyone else you think may
be interested. Please send any tips or feedback to <asb at asbradbury.org>, or
@llvmweekly or @asbradbury on Twitter.

## News and articles from around the web

The next LLVM Developers' Meeting is
[scheduled](https://twitter.com/llvmorg/status/1127938834069508097) for
October 22-23rd in San Jose.

An LLVM Social will [take place in
Chicago](http://lists.llvm.org/pipermail/llvm-dev/2019-May/132227.html) at a
yet-to-be determined date. Use the linked Doodle poll or meetup page to
indicate your preference.

GDB 8.3 [has been

## On the mailing lists

* Alexey Sidorin [provides an
update](http://lists.llvm.org/pipermail/cfe-dev/2019-May/062285.html) on
support for a matcher-like checker API.

* Praveen Velliengiri is a GSoC student working on speculative compilation in
ORC, and introduces themselves
[here](http://lists.llvm.org/pipermail/llvm-dev/2019-May/132348.html). Seiya
Nuta [will be working on improving LLVM binary

* Paul Robinson
[suggests](http://lists.llvm.org/pipermail/llvm-dev/2019-May/132298.html) the
removal of `REQUIRES: no*` instances in LLVM codebases as they may mislead
peope into thinking the `no*` prefix is understood by lit, and these cases can
be better handled using `UNSUPPORTED: `.

* Simon Pilgrim is [seeking
feedback](http://lists.llvm.org/pipermail/llvm-dev/2019-May/132196.html) on
the current scan-build reports on LLVM and interest in fixing the reported
issues. Artem Dergacehv
[indicates](http://lists.llvm.org/pipermail/llvm-dev/2019-May/132269.html) a
GSoC project will be looking at static analyzer false positives.

## LLVM commits

* LoopSimplify learned to preserve MemorySSA.

* Support was added for `AVX512_BF16` instructions, supported by the upcoming
Cooper Lake microarchitecture. [r360017](https://reviews.llvm.org/rL360017).

* On the ARM backend, register copies are now glued to tail calls. As
described in the commit message, the previous logic wasn't quite correct. If
your out-of-tree backend copied ARM you may want to review.

* InstPrinter was moved to MCTargetDesc for various backends for consistency
with backends where this is necessary in order to avoid a circular dependency.
[r360488](https://reviews.llvm.org/rL360488), and more.

## Clang commits

* P1009R2 from C++20 was implemented, allowing the array bound in an array new
expression to be omitted. [r360006](https://reviews.llvm.org/rL360006).

* C++20 explicit(bool) support was committed.

* The `-gz` and `-wA,--compress-debug-sections` options were updated to use
gABI compression rather than zlib-gnu.

* clangd gained an intermediate representation of formatted text.

## Other project commits

* The standalone version of the scudo allocator gained a Quarantine
implementation to hold chunks of memory prior to releasing them for potential
reuse. [r360163](https://reviews.llvm.org/rL360163).

* The LLDB python reference page was ported over from the old LLDB website.

More information about the llvm-dev mailing list